I think that in almost all cases, the 'professional' design is going to perform better -- where I think you'll find the conflict in advice is whether or not it's worth it to spend the expense on an untested idea.
If the idea holds enough water, it doesn't NEED a good design (re: Craigslist, tarsnap, the original Digg, etc.) -- it just needs to work. If it works, and you can prove there's a business there, then by all means, spend the money to make it perform as well as you possibly can. Design in this sense is an 'optimization' technique, not part of the core product.
The flip side, is if you have a horrible idea, and you spend $10k on design, you've effectively thrown that money away.
The exceptions to this, of course, are in the demographic. If your target market is web designers, they're going to be hypercritical of any design, so it needs to excel from day 1. If your startup is an iPhone app, same thing. Otherwise, most people will be forgiving of a spartan design, so long as the damn thing works, and that'll give you leverage to design against.
Also, another big point in this is that as you build the product, you'll change how things work, a lot. In the application I'm building now, I had Photoshop mockups for every page, and how I thought they would look, but in the initial testing, I've changed almost every page somewhat dramatically. It costs me time, which I hate, but it's better than costing me time AND dollars I'd have to pay a professional designer to re-design because I changed the position of button x. And, while I hate having to re-implement, I'd much rather not move into production with button x in the wrong place because I couldn't afford to pay a designer to move it.
If the idea holds enough water, it doesn't NEED a good design (re: Craigslist, tarsnap, the original Digg, etc.) -- it just needs to work. If it works, and you can prove there's a business there, then by all means, spend the money to make it perform as well as you possibly can. Design in this sense is an 'optimization' technique, not part of the core product.
The flip side, is if you have a horrible idea, and you spend $10k on design, you've effectively thrown that money away.
The exceptions to this, of course, are in the demographic. If your target market is web designers, they're going to be hypercritical of any design, so it needs to excel from day 1. If your startup is an iPhone app, same thing. Otherwise, most people will be forgiving of a spartan design, so long as the damn thing works, and that'll give you leverage to design against.
Also, another big point in this is that as you build the product, you'll change how things work, a lot. In the application I'm building now, I had Photoshop mockups for every page, and how I thought they would look, but in the initial testing, I've changed almost every page somewhat dramatically. It costs me time, which I hate, but it's better than costing me time AND dollars I'd have to pay a professional designer to re-design because I changed the position of button x. And, while I hate having to re-implement, I'd much rather not move into production with button x in the wrong place because I couldn't afford to pay a designer to move it.